Report No. CDOT-DTD-94-6
Implementation of a Fine Aggregate Angularity Test
Tim Aschenbrener Colorado Department of Transportation 4201 East Arkansas Avenue Denver, Colorado 80222
Final Report April 1994
Prepared in cooperation with the U.S. Department of Transportation Federal Highway Administration
The contents of this report reflect the views of
the author who is responsible for the- facts and
the accuracy of the data presented herein. The
contents do not necessarily reflect the official
views of the colorado Department of Transportation
or the Federal Highway Administration. This report
does not constitute a standard, specification, or
regulation.
i
Acknowledgements
The author would like to express his gratitude to the 19 asphalt paving contractors that participated in this study by providing the aggregates for testing. The laboratory testing was performed by Gray Currier and Cindi Moya (COOT-Staff Materials).
The COOT Research Panel provided many excellent comments and suggestions 'for the study; it included Byron Lord and Kevin Stuart (FHWA-Turner Fairbank Highway Research Center), Doyt Bolling (FHWA-Region 8), Jerry Cloud (FHWA-Colorado Division), Steve Horton and Bob LaForce (COOT-Staff Materials), Ken Wood (COOT-Region 4 Materials), and Donna Harmelink (CDOTResearch). Scott Shuler (CAPA) also provided many valuable comments.
ii
Technical Report Documentation Page
1. Heport No. 2. Government Accession No. 3. Recipient's Catalog No.
CDOT-DTD-R-94-6
4 ... iUe nod Subtitle 5_ Report Date
Implementation of a Fine Aggregate Angularity Test AjlTii 1994
6. Perfonnlng Organization Code
7_ Author(s) 8. Perronnlng Organlzatton Rpt.No.
Tim Aschenbrener CDOT-DTD-R-94-6
9. PcrrOl,nillg Organization Name and Address 10. WOI'k Unit No. (TRAlS)
Colorado Department of Transportation
4201 E. Arkansas Ave. 11. Contract or Grant No.
D"nver, CO. 80222
12. Sponsoring Agency Name and Address 13. Type of RpL and Period Covered
Cclorado Department of Transportation Final Reoort
4201 E. Arkansas Ave. 14. Sponsoring Agency Code
Denver, CO. 80222
J5. SupplcllIcntm'y Notcs
Prepared in Cooperation with the U.S. Department of Transportation Federal
Highway Administration
16. Abstract
Both the contractors and Colorado Department of Transportation have had frustrations with the specification used to limit
natural fine aggregates. The angularity of manufactured and natural fine aggregates needs to be more clearly defined for
c\,Jlltr(tc~u aJ purposes. TIle purpose of this report is to develop a specification and implementation plan for quantitatively
defining manufactured and natura] fine aggregates.
Two methods were investigated: the AASHTO TP3 and French P18-S64 methods. Re'mlts from the two tests provided
L'OmparJbJc rankings of the individual stockpiles nnd blended natural and manufactured fme aggregates tested in this study.
'lhe AASI ITO 11)3 is recommended because of snmple preparation time. Specifications to limit the natural tine aggregates
in quan:itative manncr and an implementation plan were,; developed.
17. Kc)' Words 18. Distribution Statement
Fine Aggregates NAA Flow Test No Restrictions: This report is
Angularity Natural Sands available to the public through
the National Technical Info.
Service. Springfield, VA 22161
19.5cc:urlty Classif. (report) 20.Secllrlty Classlf. (paGe) 21. No. of Pages 22. Price
U nclassi fied Unclassified 61
iii
Table of Contents
1.0 Introduction .............. .. ' .......... . ... ... .... . . . . .... . .... . . . .. . 1
2.0 Fine Aggregate Angularity Tests ..... .... . ... . ... . . . . ..... . .. , . . . . . . . .. 2 2.1 AASHTO TP3 (The National Aggregate Association Test) . . .. . . . .. . . . . . . 2 2.2 French P 18·564 ........ ..... .. . .. ..... . . . . . .. . .. . .. , . . . . . • .. 4
2.2.1 The French Method ... ... .. . . . . . . .. . . . .. . . . . . . . . . . . . . . . 4 2.2.2 Modifications for This Study ....• . ... . ..... . . • . . . . . . . . . • .. 6
3.0 Previous Studies and Recommended Specifications. . . . . . . . . . . . . . . . • . . . . . • . . 7 3.1 Kandhal (5) .. . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3.2 Mogawer and Stuart (6) . . . . . . . . • . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . 7 3.3 Aschenbrener (7) .. . ... . . . . .... . .. . . • . .. . ... . . .. . . ... . . . . ... . 8 3.4 SHRP . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 8
4.0 Test Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . • . . . .. 10 4.1 AASHTO TP3 .... . .... . . . ........ . ... . . .• . .. . .. . .. .. . .. . . . . . 10
4.1.1 Individual Stockpiles . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . 10 4.1.2 Combined Blends . .... . . . . . .• . . .. .. . ... , . . .... . . .. ... . 10 4.1.3 Calculating vs. Testing Combined Blends ...... . ...•... . . • . . . 14
4.2 French P 18-564 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 4.2.1 Individual Stockpiles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . 15 4.2.2 Combined Blends ...... ....... ... . ... . . .. . .. . . . .. . . . . . 15 4.2.3 Calculating vs. Testing Combined Blends . . . . . . ... .... . . . . .. . 18
4.3 Comparison of the AASHTO TP3 and French P 18-564 Methods . . . . . . . . .. 18 4.3.1 Correlation of Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 4.3.2 Time of Preparation .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 4.3.3 Other Considerations .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.4 Recommended Specification Values . • . . . . . . . . . . . . . • . . . . . . . . . . . . . .. 21
5.0 Repeatability . .. .•.. . ..... . ... . ... .. .. .. . . ..... .. . ... . . . . . • . ...... 23
6.0 Implementation . .............. ..... . .. . . .. . . . .. ... . .... . . . , . .. . ... 24 6.1 Implementation Plan for Projects . .. . .. . .... . . .. . . . .. . . • . . , . .. . ... 24
6.1.1 1994 Construction Season . . .. .. ......... . .. . . . . .. .. .. . . . 24 6.1.2 1995 Construction Season . . ......... . . . . . . . .. . . . . . . • . ... 24
6.2 Draft Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 6.3 Other Modifications Necessary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
7.0 Concl usions . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 26
iv
Appendices
Appendix A: AASHTO TP3 Appendix B: French Tentative Standard P 18-564
List of Tables
Table 1. Summary of Previous Studies Using AASHTO TP3, Method A. .. . . . . . . . . . .. 7 Table 2. Minimum Uncompacted Voids ('Yo) Specified by SHRP ....... ... . .. . . .. . " 9 Table 3. Summary of Fine Aggregate Angularities. . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 11 Table 4. Comparison of Measured and Calculated Results of Combined Fine
Aggregates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 14 Table 5. Correlation of AASHTO TP3 and French P 18-563 Methods. . . . . . . . . . . . . . . . 20 Table 6. Recommended Specification Values for AASHTO TP3. _ .. _ .. _ . . . . . . . . . . 21
List of Figures
Figure 1. The AASHTO TP3 Angulometer. . .... . .... . . .. . .. . .. . .... . .. . .. • .. 3 Figure 2. Schematic of the AASHTO TP3 Angulometer. . . . . . . . . . . . . . . . . . . . . . . . .. 3 Figure 3. Schematic of the French Angulometer. ..... . . . .. . .......... . . . . .. . .. 5 Figure 4. Ranked Order Plot of Uncompacted Voids ('Yo) of Individual Stockpiles Using
AASHTO TP3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 12 Figure 5. Ranked Order Plot of Uncompacted Voids ('Yo) of Blends using AASHTO
TP3 ........................................................ . . 13 Figure 6. Ranked Order Plot of Flow Time of Individual Stockpiles Using French P 18-
564. ....................................................... .. 16 Figure 7. Ranked Order Plot of Flow Time of Blends Using French P 18-564. .. . . . . . .. 17 Figure 8. Comparison of Flow Time from the French P 18-564 with the Uncompacted
Voids from AASHTO TP3. . ... ..... . . ... .. .... . . . . .. .. .. . .. . . ..... . 19
v
Implementation of a Fine Aggregate Angularity Test
Tim Aschenbrener
1.0 Introduction
Rutting of a hot mix asphalt (HMA) pavement is primarily related to aggregate structure and
secondarily related to asphalt cement stiffness. One of the key components of the aggregate
structure is the angularity of the fine aggregate. The angularity is oiten referred to as the particle
shape and surface texture.
The Federal Highway Administration's (FHWA's) Technical Advisory T 5040.27 (1) recommends
limiting the natural fine aggregates by the weight of the total aggregate blend to 15% through 25%
dej:ending on traffic. Since the 1990 construction season, the Colorado Department of
Transportation (COOT) has limited natl,Jral fine aggregates to a maximum of 20%.
Although limiting natural fine aggregates is believed to have improved the performance of the
HMA pavements in Colorado, the specification has been contested. First, it is not easy to define
a natural fine aggregate. Some fine aggregates are close to their parent rock and have not
weathered; the natural fine aggregate is actually quite angular. In other instances, contractors
have placed large quantities of fine aggregates through a crusher that must have been turned off.
The 'manufactured" fine aggregates appear no more angular than the natural fine aggregates.
Second, not all manufactured fine aggregates have the same angularity. When using highly
angular manufactured fine aggregates, perhaps 20% natural fine aggregates would be acceptable.
HOlVever, when using the marginal, less-angular fine aggregates, perhaps 0% natural fine
aggregates should be allowed. Currently, the blend of the manufactured and natural fine
aggregates is not considered.
Both the contractors and COOT have had frustrations with the specification . The angularity of
manufactured and natural fine aggregates needs to be more clearly defined for contractual
purposes. The purpose of this report is to develop an implementation plan and specification for
quc.ntitatively defining manufactured and natural fine aggregates.
1
2.0 Fine Aggregate Angularity Tests
Numerous tests exist to measure the particle shape and surface texture of fine aggregates.
Some of the tests include:
1) ASTM D 3398, Standard Test Method for Index of Aggregate Particle Shape and Texture; 2) ASTM D 3080, Direct Shear Test; 3) the National Aggregate Association's Methods A, B, and C, (now referred to as AASHTO TP3); 4) the National Sand and Gravel Association Test (2); 5) Michigan's fine aggregate angularity test (3); and 6) the French P 18-564, Determining the Flow Coefficient of Sand (4).
Only the AASHTO TP3, Method A, and the French P 18-564 methods were investigated for
implementation.
2.1 AASHTO TP3 (The National Aggregate Association Test)
The National Aggregate Association (NAA) has developed a test to objectively quantify the
angularity of fine aggregates. The NAA test was used to develop the draft AASHTO TP3 entitled
"Standard Test Method for Uncompacted Void Content of Fine Aggregate (As Influenced by
Particle Shape, Surface Texture, and Grading)".· The test procedure is included in Appendix A.
The angulometer is shown in Figure 1 and a schematic is shown in Figure 2.
Fine aggregates flow freely into a 1 OO-mL copper cylinder. By knowing the bulk specific gravity
of the fine aggregate, G'b' and the weight of fine aggregate in the cylinder of known volume, the
uncompacted void content can be calculated. Very angular fine aggregates will have high
uncompacted voids, and more rounded fine aggregates will have low uncompacted voids.
The draft AASHTO TP3 includes Methods A, B, and C. Method A was LJsed for this study. In
Method A, a standard gradation is always tested by combining a pre-determined quantity of
individual sieve fractions from a typical fine aggregate sieve analysis (AASHTO T 27). In Method
B, individual sieve sizes are tested separately. In Method C, the "as-received" gradation is
tested.
2
Figure 1. The AASHTO TP3 Angulometer.
Pan to Retain Fine Aggregate
Particles
I ! ,
An8 Quart (Uter)
Mason Jar Pycnometer (Remove botto.m of glass Jar)
i W±4·
/iT 12.1±O.6 mm dla I'
11S±2mm !
Nominal 100 mL Meksure I ,,: '--L i
Section Through Center of Apparatus
Figure 2. Schematic of the AASHTO TP3 Angulometer.
3
The size of sample required for the test is always 190 grams. In Method A, the size of the fine
aggregate tested is between the 2.36 mm (No.8) sieve and the 0.150 mm (No.1 00) sieve. In
Method B, the size of the fine aggregate tested is between the 2.36 mm (No.8) sieve and the
0.300 mm (No. 50) sieve. In Method C, the fine aggregate tested is passing the 4.75 mm (No.
4) sieve.
The uncompacted voids are measured for 2 determinations. The result is the average of the 2
determinations reported to the nearest tenth.
Me!hod A was used for this study. Method A is preferred over Method B because of preparation
time. All of the fine aggregate can be determined by preparing one sample. Three samples,
each of different sizes, must be prepared to use Method B. Method A is preferred over Method
C because the influence of gradation on the uncompacted. voids is not a factor. Method A
requires one standard gradation for all tests.
2.2 French P 18-564
2.2.1 The French Method
The French Method P 18-564 entitled "Determining the Flow Coefficient of Sand" is included in
Appendix B. A schematic of the angulometer is shown in Figure 3. Aggregates flow freely
thrcugh an orifice. The time required for 1000 grams of fine aggregate to flow through the orifice
is recorded. The faster the fine aggregates flow through the orifice, the less angular the material.
The "as-received" gradation is tested. Two different sizes of samples are allowed. The fine
aggregates passing the 4.75 mm (No.4) sieve are tested with an orifice that has a 16-mm
diameter opening, or the fine aggregates passing the 2.36 mm (No.8) sieve are tested with an
orifice that has a 12-mm diameter opening. Five determinations of flow time are made, and the
time is recorded to the nearest tenth of a second for each determination. The final result is the
average time reported to the nearest second.
4
p 18 - 50!.
'-.- ' '-...: , .x. --<. /---
"- ~ 90
" in '1--.--=+.l- - - - - - - - - --- - . .1; " +--
.r ~~4-U. 1 ; Ii r _____ I" mi.ll i _t res
Funne.l
t Co ve r pl.lce
Con.t .& in. e c
Figure 3. SchematiC of the French Angulometer. I
5
i , ,
(!
/'J /
i , ., , . \ \
, . ", \
The French have no specifications for the results. The results are only used an indication of the
fine aggregates ability to resist rutting. If the flow time is less than 35 to 37 seconds then there
exists a risk that there will be a problem with rutting . When the flow time is short, the French
rutting tester is used to provide the final acceptance of the proposed HMA.
2.2.2 Modifications for This Study
The "as-received" gradation was not used forth is study. Since changes in gradation are known
to influence the uncompacted void content, samples were tested using one standard gradation.
The gradation used was selected to be the same as recommended in AASHTO TP3. The entire
fine aggregate passed the 2.36 mm (No.8) sieve and was retained on the 0.150 mm (No.1 00)
sieve. The quantity of material tested remained at 1000 grams as stated in the procedure.
As stated in the procedure, five determinations were taken for each fine aggregate and the time
was recorded to the nearest tenth of a second. These measurements were remarkably similar.
For this study the average of the five determinations was reported to the nearest tenth of a
second, even though it is normally reported to the nearest whole second.
6
3.0 Previous Studies and Recommended Specifications
Pre'fious studies by Kandhal (5), Mogawer (6), and Aschenbrener (7) have investigated the use
of the NAA fine aggregate angularity test (now referred to as AASHTO TP3) with natural and
manufactured fine aggregates. The results are summarized in Table 1. The most rounded and
most angular fine aggregate tested in each study is listed in Table 1. Additionally, each author's
recommended cutoff to delineate between natural and manufactured fine aggregates is included
in Table 1. These studies all reported test results from single aggregate stockpiles; blends of
natural and manufactured fine aggregates were not tested.
Table 1. Summary of Previous Studies Using AASHTO TP3, Method A.
I Uncompacted Voids (%)
Kandhal (5) Mogawer (6) Aschenbrener (7)
Most Rounded 40.2 40.1 41.2
Most Angular 52.0 50.7 50.4
Cutoff Between 44.5 43.4 44.0 Round vs. Angular
3.1 Kandhal (5)
A study was performed on 18 fine aggregates from various sources throughout Pennsylvania.
The natural fine aggregates had an average uncompacted voids of 42.5%. The manufactured
fine aggregates had an average uncompacted voids of 48.1%. An uncompacted void level of
44.5% differentiated between the angularity of manufactured and natural fine aggregates.
The AASHTO TP3, Method A, was found to correlate very well with the accurate but time
consuming ASTM D 3398.
3.2 Mogawer and Stuart (6)
Tests were performed on 12 fine aggregates from Maryland, New Jersey, Virginia, and Wisconsin.
Using the AASHTO TP3, Method A, the "poor quality natural sands' had an average
uncompacted voids of 41.7%, the "good quality natural sands' had 45.6%, and the "manufactured
7
sands" had 47.2%. An uncompacted void level of 43.4% separated the good and poor quality
natural fine aggregates.
The AASHTO TP3, Method A, was found to correlate very well with the accurate but time
consuming ASTM D 3398.
3.3 Aschenbrener (7)
A study was performed on 25 fine aggregates from the 17 most commonly used aggregate
sources in Colorado. Manufactured fine aggregates were tested from 6 quarries and 11 sand
and gravel pits; natural fine aggregates were tested from 8 of the sand and gravel pits. Fine
aggregates obtained from quarries had an average uncompacted void content of 49.2%. Fine
aggregates manufactured from sand and gravel pits had an average uncompacted void content
of 46.6%. The natural fine aggregates from sand and gravel pits had an average uncompacted
void content of 43.4%. The cutoff used to distinguish between natural and crushed fine
. aggregates was 44.0%.
It was interesting to note that three of the eight natural fine aggregates had angularities greater
than the cutoff. These sands are close to their parent rock and have not weathered extensively.
On the other hand, two of the "manufactured" fine aggregates were less than the cutoff. These
"manufactured" fine aggregates were not processed in a quality manner.
3.4 SHRP
SHRP has recommended the use of AASHTO TP3, Method A. The specified values should be
applied to the blend of fine aggregates proposed for use in HMA. SHRP specified values of
uncompacted voids (8) as a function of traffic and depth of the pavement layer from the surface
as shown in Table 2. Traffic is defined as the traffic in the design lane for the design life. For
roads with less than 3 million design ESALs, it appears that 100% natural fine aggregates are
acceptable for surface mixtures.
8
Table 2. Minimum Uncompacted Voids (%) Specified by SHRP.
Uncompacted Voids (%) Traffic,
Depth from Surface ESALs
< 100 mm > 100 mm
< 3 x 105 - .
< 1 X 10· 40 -
< 3 X 10· 40 40
< 1 X 107 45 40
< 3 x 107 45 40
< 1 X 10· 45 45
>3 x 10· 45 45
Although the SHRP recommendations appear very low, it should be noted that there are two
significant differences in the SHRP specification. The SHRP gradation is coarser than the CDOT
Master Range and has a restricted zone. The coarser gradation allows for a smaller percentage
of fine aggregate. The restricted zone may prevent the use of some natural fine aggregates.
Additionally SHRP has a minimum requirement for the air voids at the initial number of gyrations,
N,nN, on the SHRP Gyratory Compactor. These two differences might limit the natural fine
aggregates further than the angularity requirement.
9
4.0 Test Results and Discussion
Contractors from around Colorado submitted 19 HMA mixtures that were commonly used on
CDOT projects. These mixtures were designed using variable laboratory compactive efforts and
tested for strength using the Hveem .stabilometer and the French rutting tester (9) . The summary
of test results on the fine aggregates using the AASHTO TP3 and the French P 18-564 methods
is shown in Table 3.
In some cases, the contractors submitted HMA mixtures with no natural fine aggregate. These
are indicated with "None" in the "Nat." column of Table 3. In some cases, we had a shortage
of material and were unable to perform the fine aggregate angularity tests. These are indicated
with "NT", not tested, in Table 3.
4.1 AASHTO TP3
4. 1. 1 Individual Stockpiles
The uncompacted voids have been plotted in ranked order in Figure 4. Some of the sites are
. listed twice because the contractor used a natural fine aggregate stockpile and a manufactured
fine aggregate stockpile. The delineation between natural and manufactured fine aggregates
appears to be at an uncompacted void level of 45.0%. There are 3 natural fine aggregates that
have higher uncompacted voids than 45.0%. In each of these 3 cases, the material is a washed
concrete sand. An uncompacted voids of 46.5% would exclude aI/ of the natural fine aggregates,
except one (Mix 3) .
4. 1.2 Combined Blends
The percentage of manufactured and natural fine aggregates in the total blend of aggregates
passing the 2.36 mm (No.8) sieve was calculated and is shown in Table 3. The manufactured
and natural fine aggregates were then blended using these proportions and tested using AASHTO
TP3. A ranked order plot of the uncompacted voids of the combined blends of fine aggregate
is shown in Figure 5. In aI/ cases, the combined fine aggregates had an uncompacted void level
greater than 44.0%.
10
Table 3. Summary of Fine Aggregate Angularities.
Percentage (%) AASHTO TP3 French P 18-564 Passing the 2.36
Uncompacted Voids (%) Time (seconds) Mix mm Sieve
Man. Nat. Man. Nat. Comb. Man. Nat. Comb.
1 65 35 47.6 44.3
2 75 25 46.8 NT
3 65 35 49.7 46.6
4 100 0 48.2 None
5 65 35 48.8 45.8
6 40 . 60 48.8 44.4
7 75 25 47.0 42.7
8 65 35 48.6 41.7
9 50 50 45.1 44.5
10 100 0 46.8 None
I 11 70 30 49.7 44.1
! 12 70 30 47.7 43.3
13 100 0 48.6 None
14 · 80 20 46.6 44.0
I 15 70 30 48.1 43.9 I
I 16 65 35 47.0 45.8 I
i 17 100 0 48.3 None
Ii Il 18 85 15 47.0 43.5
II 19 100 0 47.3 None
Nat. • Natural Man. - Manufactured Comb.- Combined Natural and Manufactured
11
NT
NT
48.5
48.2
NT
46.1
NT
46.6
44.0
46.8
48.3
46.9
48.6
46.9
46.9
46.0
48.3
46.2
47.3
NT NT NT
35.1 NT NT
38.1 33.3 35.9
36.2 None 36.2
37.0 34.2 NT
37.3 31 .1 32.5
34.0 30.0 NT
39.3 28.8 34.5
35.4 32.9 32.6
37.1 None 37.1
38.3 31.7 34.3
36.7 30.1 34.7
39.6 None 39.6
35.9 31 .2 34.1
37.0 31 .3 34.8
35.4 32.7 33.4
38.0 None 38.0
35.9 31 .9 34.4
33.6 None 33.6
None - No Natural Used in HMA NT - Not Tested
-t-Il)
'"C ·0 > '"C
50.0~~~~~~~~,~! ~!~I~I~i-I~I~!~I~I~~i~I~I~i~! ~I~l~l~-*~: ~*~
49.0
i 45.0~-4-4-4-4-4-4-4-4i_~~!-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+i~-+-+-+-H ~ i i ~ ~ i ! ! ! ! ! 8 44.0 c
;:)
::::~···rr···r····:····r··r·rr··!··r····r·r·l·····i··r··r·r··r··r··r····i·--r··r·r····i··;--·rr··r 8 71218151411 6 9 9165 31421071618191215417138 6 5 311
Site
D Natural *' Manufactured
Figure 4. Ranked Order Plot of Uncompacted Voids (%) of Individual Stockpiles Using
AASHTO TP3.
12
49.0
6 ~ 46.04-+---e-~¥---~~---+---+--~--__ --~~+---~--~~---+~1
I 1
45.0 .. _-_.+ .............. ! .. ·--·--·······f·-----·.·.····t·······.·--·--l·········· ...... l.i,· __ ···········l········· ····t········· __ ···t-·············!··············+··········· __ ·t··············f··· __ ··· __ ···+········ __ · __ ·j ...... . : 1 f ' ' !
i 44.0-
41.01-L~-----~~---,~---,+i ----,4----,,~i----~~~---,~i----,~---+-T---~T. --~i~---'~ I ___ ,,+i ____ ,,~i~ 9 16 6 18 8 10 15 12 14 19 4 11 17 3 13
Site
Figure 5. Ranked Order Plot of Uncompacted Voids (%) of Blends using AASHTO TP3.
13
Mix 9 was very unusual. The combined fine aggregate angularity was lower than either of its
components. This occurred for both the AASHTO TP3 and the French P 18-564 methods.
When Mix 9 is excluded, all of the uncompacted voids for the combined natural and manufactured
fine aggregates equalled or exceeded 46.0%.
4.1.3 Calculating vs. Testing Combined Blends
The uncompacted voids of each individual stockpile is known. Since the percent of each
individual stockpile is known, it is possible to calculate the uncompacted voids of the blended fine
aggregates. ThE;! comparison of the measured and calculated uncompacted voids of the blended
fine aggregates is shown in Table 4. The comparison is made for 10 of the 19 mixtures
investigated. Ablend was not compared if the combination was not tested or there were no
natural fine aggregates.
Table 4. Comparison of Measured and Calculated Results of Combined Fine Aggregates.
AASHTO TP3 French P 18-564
Mix Uncompacted Voids (%) Time (seconds)
Measured Calculated Ditt. Measured Calculated Ditt.
3 48.5 48.6 +0.1 35.9 36.4 +0.5
6 46.1 46.2 +0.1 32.5 33.6 + 1.1
8 46.6 46.2 -0.4 34.5 35.6 · + 1.1
9 44.0 44.8 +0.8 32.6 34.2 +1.6
11 48.3 48.0 -0.3 34.3 36.3 +2.0
12 46.9 46.4 -0.5 34.7 34.7 0.0
14 46.9 46.1 -0.8 34.1 35.0 +0.9
15 46.9 46.8 -0.1 34.8 35.3 +0.5
16 46.0 46.6 +0.6 33.4 34.5 + 1.1
18 46.2 46.9 +0.7 34.4 . 35.3 +0.9
~ +0.02 +0.97
S.D. 0.54 0.54
Negative numbers indicate the calculated is lower than the measured.
14
For the 10 samples tested. the average difference between the .measured and calculated values
was negligible. However. the calculated values were as much as 0.8 higher than the measured
values (Mix 9) and as much as 0.8 lower than the measured values (Mix 14). The standard
deviation of the differences was 0.54. There seems to be considerable variability of the
uncompacted voids between measured and calculated values. It is not recommended to use an
uncompacted voids calculated from individual components. However. If all individual stockpiles
exceed the specified value. it would probably not be necessary to test the blended fine
aggregates.
4.2 French P 18-564
4.2. t Individual Stockpiles
The flow times have been plotted in ranked order in Figure 6. Some of the sites are listed twice
because the contractor used a natural fine aggregate stockpile and a manufactured fine
aggregate stockpile. The delineation between natural and manufactured fine aggregates appears
to be lower than the French recommendation. The delineation is around 33.0 to 35.0 seconds.
and the French recommend 35 to 37 seconds. There are two manufactured and two natural fine
aggregates in with results between 33.0 and 35.0 seconds.
4.2.2 Combined Blends
The percentage of manufactured and natural fine aggregates in the total blend of aggregates
passing the 2.36 mm (No.8) sieve was calculated and is shown in Table 3. The manufaCtured
and natural fine aggregates were then blended using these proportions and tested using the
French P 18-564.
A ranked order plot of the flow time of the combined blends of fine aggregate is shown in Figure
7. Only two of the fine aggregate blends (Mixes 6 and 9) were shorter than 33.0 seconds. All
of the other blends exceeded 33.0 seconds. It was interesting to note that only 5 of the 15
blends tested greater than 35.0 seconds. Eight of the blended fine aggregates tested were in
the range of 33.0 to 35.0 seconds.
15
36.0- ·..f.··f···-+-
34.0~ +··-t·-·t·r····i······t·-·t-·t·"·'r--I"-1--l-···~ ... 9--+-- I'-+"+-'1"I-····I·····L+-·j·-I· I-·-t·· ···t'-1- .. 1 ..
i. .+ ..... , ... ~ .. rb ... ~ ... '.~"'f .+.i······l····+· ··i .... f ..... + ... .. ,i.':" '.,l ... '~"""'f"" "'~""'+"-'.j ....... I.t···i. i ...... t .. ,l ... ,( .... ..f. .. ~ ~ 9 : : i j ) ~
28.0--
26.0~,-+-+--r-+-+--r-+-+-,-r-+-+~--r-+-4--r-+-+~~+-+-, -r-+-+~--~, +-4-~ 8 712614151118169 3197 5 2 916181441251510617 3 11 813
Site
o Natural Manufactured
Figure 6. Ranked Order Plot of Flow Time of Individual Stockpiles Using Fren.ch P 18-564.
16
o "'C I: o U Q) II) -Q)
E j::
- l i ! !
38.0- ····+-····· ·; .. · .... ···-!-· ····· ·+ .. · ...... I ...... ··+ .. i , ii' i i Ei:J i j j ~ j i ·······)·············_-.... ".r'············ r·············· j ············-'r···--· __ ····· r············-'1 ... ---------- · .... ,f ............. j _._-_ ..
ii i i Ul
3··, t··+-+·-j +--f+ --r+-j---tr-+j- -j· , ! , rb 9
34.0- ····..1.····· ...... .:·· ·····:., .. : .. ·· .... 1 ... , ...... ~., ..... ,' ....... ... ?, ..... ,~ . ..... i ........... t ..... . L ........ ...! . ..... + ... ... . ~ .......... ! ........ +.. ... ! ... ~ LW ! i ! i I I ! [ .
•. J, o ¢ " ., i 32 0- ._ .. .l ... ........... l. ..... _ ....... i ............ -l .......... -.. !- ... --- .... J ............. J. .. -........ i ........... -.l .. ----...... i .......... -.-4-... - -- -- . .;. .............. i._ ... ____ __ . .i. __ ......•••... i.. __ .
. I t r iii r i . . : : : : . :
- i i !' I 30.0- .... i .. · .. _·-t -.... ····i-·_·--+· ........ ·!· .. ·--t---l .. · .. ·-·-r ...... r .. _·-r--·1 ...... .. r -·-/·--.. ··r··_ .. r· .. ·
i i. l: i [ i! . 28.0- ..... l· .. · .. ··T· · .. · ...... !· .... ·_j ·· .... · .... I .... ·· .. + ...... · .... i ......... + ....... j ........ + .. ...... f ............ j.. .. .. + ....... +. ... -l- .. ..
iii 26.0~~----t---~----~,---,~--~,r---~---~,---+----t---~~--~---~,--~--~,~
6 9 16 19 14 11 18 8 12 15 3 4 10 17 13
Site
Figure 7. Ranked Order Plot of Flow Time of Blends Using French P 18-564.
17
4.2.3 Calculating vs. Testing Combined Blends
The uncompacted voids of each individual stockpile is known. Since the percent of each
indi'/idual stockpile is known, it is possible to calculate the uncompacted voids of the blended fine
aggregates. The comparison of the measured and calculated uncompacted voids of the blended
fine aggregates is shown in Table 4.
For the 10 samples tested, the calculated times averaged nearly 1 second longer than the
measured values. However, the calculated values were as much as 2 seconds longer than the
measured values (Mix 11) and were never shorter than the measured values. The standard
deviation of the differences was 0.54. There seems to be considerable variability of the time
between measured and ca:lculated values. It is not recommended to use a time calculated from
individual stockpiles. However, if all individual stockpiles exceed the specified value, it would
probably not necessary to test the blended fine aggregates.
4.3 Comparison of the AASHTO TP3 and French P 18-564 Methods
4.3.1 Correlation of Test Results
As the uncompacted voids measured by AASHTO TP3 increases, the flow time in the French P
18-564 also increases as shown in Figure 8. The equation of the regression line is:
where:
T = 1.27 U, - 24.1
Us = uncompacted voids (%) from AASHTO TP3, Method A, and
T = flow time (seconds) from the French P 18-564.
The coefficient of determination, r, was 0.86. The correlation of several uncompacted voids to
flow times are shown in Table 5. An uncompacted voids of 45.0% from AASHTO TP3 is
approximately equivalent to 33.0 seconds using the French P 18-564. When the flow time is
greater than 35.0 seconds, the French have a lot of confidence the fine aggregate angUlarity. A
flow time of 35.0 seconds is approximately equivalent to an uncompacted voids of 46.5%.
18
40.0.-----~.----~,~----~,------r--lp-a-s-s~~-o-t-h-I--'. ----~:----~--'j-----,
38.0 ··· __ ······· ... ········--i························f··· .. ·· .. ················1--······················ ······· __ ···· __ ··········f .. ············· __ · __ ·····+········ __ ··········--·f·····* · __ · __ ····· ··1-··············.1.·····
r 2= oj86 j i ')I( : lI': ~ ! ····· .......... f .. ·· ······ .. r· .. · ·· .. r.. · .. ·· · ~·.. .. · · j ........ *=·~.. · .... ··l~· .. .... · .. ·i·· .. ·· .. .... ···
!! Ii . 36.0
'iii" "C I: o 34.0 .... · .. ·...... t .. ............ ·· ·t .. ........ · .... ·t .. · ...... · .. · .. · .. · .. · ...... ( ...... ~· .... ~t~· .... · .. +· .... · ...... ··j .. · .... ·.... .. u G)
Cl -! ' I/) -G) 32.0 .. ····· ...... ·r· .. · .. · .. ··· .. ·r .... · .. Cl'··~ .. ·:·· · ......... . .... 1""""""" ""1""""""""t· .. · .. · .. · .. ··, ...... · ...... · .... .. E ;:: . 1 1 i l
30.0 · -.--.~-.... o ......... -.-.-.i-..................... -............. · .. · ......... i ............... ······ .... ! ........................ .;..-...................... ! .......... _ ............ . : . , .: i i
28.0 ·················_·_·f························i-······ ................ +_ ..................... -.-... -.--............ ;.-----------------...... f •••••• ----------------··t···········--------------i------··· ......... -----.-
26.0+-----4!----_4i--~I=Fa=i~I~=O=t=h~I~----_4:--~_4----_4·~--~!,_--~ 41 42 43 44 45 46 47
Uncompacted Voids (%)
o Natural *' Manufactured
48 49
Figure 8. Comparison of Flow Time from the French P 18-564 with the Uncompacted Voids
from AASHTO TP3.
19
50
Table 5. Correlation of AASHTO TP3 and French P 18-563 Methods.
Uncompacted Corresponding Voids (%) Flow Time
(Seconds)
45.0 33.0
46.0 34.3
46.5 35.0
48.1 37.0
4.3.2 Time of Preparation
In order to prepare a sample for testing, a standard gradation (AASHTO T 27) is performed. The
fine aggregates on the sieves used for the angularity test are saved. When a standard gradation
(AASHTO T 27) is performed with 1000 grams, it typically takes 1 gradation to obtain enough
material to perform AASHTO TP3, Method A, with 190 grams. AASHTO TP3, Method B, may
take more than 1 gradation. However, it typically takes 3 gradations to obtain enough material
to perform the French P 18-564 with 1000 grams. Approximately 30 additional minutes are
required to prepare a sample lor testing using the French P 18-564 method than the AASHTO
TP3 method.
4.3.3 Other Considerations
When testing with the AASHTO TP3 method, the bulk specific gravity of the fine aggregates, G.b' as measured by AASHTO T 84 is required. AASHTO T 84 is always performed for the mix
design to calculate the VMA of the compacted HMA mixture. However, AASHTO T 84 is
periormed on all of the material passing the 4.75 mm (No. 4) sieve. AASHTO TP3 is performed
on material passing the 2.36 (No.8) sieve and retained on the 0.15 mm (No.1 00) sieve.
Furthermore, AASHTO T 84 is measured on the "as-received" gradation, and AASHTO TP3 is
performed on a standard fabricated gradation. Differences in the sizes and gradations of the fine
aggregate when testing AASHTO T 84 and AASHTO TP3 could cause for misleading angularity
results.
20
The French P 18-564 method could be influenced by the specific gravity of the aggregates. If
a light-weight aggregate were tested that was very round, it would still take a longer time to flow
through the orifice since there would be more volume of fine aggregate. If the French procedure
is used, an adjustment should be made when using fine aggregate with specific gravities
substantially different than the st.andard specific gravities. The bulk specific gravities of all of the
aggregates used In this study were between 2.562 and 2.695.
4.4 Recommended Specification Values
Currently the CDOT limits natural fine aggregates in a mixture to a maximum of 20%. It is not
desired to implement an angularity test to allow higher quantities of natural fine aggregates.
Additionally, it is not desired to impose hardships on contractors because of a new angularity
specification. The CDOT is generally pleased with the fine aggregate angularity in its current
HMA mixtures.
AASHTO TP3 is recommended for use as a specification. The MSHTO TP3 procedure was
considered better than the French P 18-564 because of the shorter preparation time. The test
should be performed on the blend of fine aggregates, and calculations should not be allowed.
The recommended specification values for the AASHTO TP3 are shown in Table 6.
Table 6. Recommended Specification Values for AASHTO TP3.
Traffic Uncompacted Voids (%) ESALs Blended Fine Aggregates
< 1 x 10· 45.0
> 1 x 10· 46.0
The uncompacted voids of 46.0% for high trafficked roads will allow all but one of the mixtures
(Mix 9) to pass. These are plotted in ranked order in Figure 5. It Is interesting to riote three
examples of mixtures that barely meet the recommended specification. These examples provide
an indication of the quantity of natural fine aggregates that will be allowed. Mix 6 utilized 35%
washed concrete sand with a manufactured fine aggregate from a quarry. The uncompacted
21
voids were 46.1 %. Mix 16 used 20% washed concrete sand with a manufactured fine
aggregates from a sand and gravel pit. The uncompacted voids were 46.0%. Mix 18 used 10%
natural fine aggregates with a manufactured fine aggregate from a sand and gravel pit. The
uncompacted voids were 46.2%.
All of the mixes tested would pass the specification of 45.0% uncompacted voids, except Mix 9.
Mix 9 was crushed from a sand and gravel pit and has 20% natural fine aggregates.
An uncompacted voids of 46.0% is approximately equivalent to a flow time of 34.3 seconds using
the French P 18-564 method. It should be noted that a flow time of 34.3 seconds would cause
concerns by the French that rutting might be a problem. An uncompacted voids of 45.0% is
approximately equal to a flow time of 33.0 seconds.
The uncompacted voids of 45.0% will allow more than 20% natural fine aggregates into an HMA
mixture. This is a potential concern and should be implemented with caution. Although an
uncompacted voids of 45.0% causes concern, SHRP has recommended (8) an uncompacted
voids of 40.0% for less than 3 million ESALs. SHRP would easily allow 100% natural fine
aggregates.
Although the SHRP recommendations appear very low, it should be noted that there are two
significant differences in the SHRP specification. The first involves the aggregate gradation.
The SHRP gradation is coarser than the CDOT Master Range and has a restricted zone. The
coarser gradation allows for a smaller percentage of fine aggregate. The restricted zone may
prevent the use of spme natural fine aggregates. Secondly SHRP has a minimum requirement
for the air voids at the initial number of gyrations, Ntnlt, on the SHRP Gyratory Compactor. These
two differences might limit the natural fine aggregates further than the angularity requirement.
The SHRP fine aggregate angularity requirements should be investigated before implementation.
22
5.0 Repeatability
AASHTO TP3 has a precision statement. The single operator standard deviation is 0.13 percent
voids. Two properly conducted tests by the same operator on similar samples should not differ
by more than 0.37 percent uncompacted voids. The multi-laboratory standard deviation is 0.33.
Two properly conducted tests by different operators on similar samples should not differ by more
than 0.93 percent uncompacted voids.
Based on testing for this study, the Single-operator standard deviation may be lower than the
precision statement. Additional studies should be performed to verify the precision statement
in the procedure.
The bulk specific gravity of the aggregate ((3.b) as determined by AASHTO T 84 is very important
in deiermining the uncompacted voids. A change in the GSb by 0.05 will change the
uncompacted voids by approximately 1.0% uncompacted voids. A change in GSb by 0.005 will
change the uncompacted voids by approximately 0.1%.
23
6.0 Implementation
6.1 Implementation Plan for Projects
6. 1.1 1994 Construction Season
Step 1:
Step 2:
Step 3:
Step 4:
Step 5:
Step 6:
The Central Laboratory has ordered seven angularity devices, one for each Region Laboratory. It is uncertain when they will be available because there is no commercial distributor of the equipment at this time.
A Colorado Laboratory Procedure (CP-L) should be developed. Each Region should receive the procedure, device and training simultaneously. A round robin should be performed to investigate within- and between-laboratory variability.
Each Region Laboratory should perform the test on projects when gradation is tested. The angularity test results should be provided to the contractors. The test results will also be useful to identify potential causes for loss of air voids when field verifying mixes.
The test results and percent of natural fine aggregates should be submitted to the Central Laboratory for compilation and analysis. A baseline of angularities can be obtained for the entire state.
At the end of the construction season, the specified values should be re-evaluated. If the season of monitoring is successful with no problems, then the specification could be considered for statewide implementation. If. problems arise, then a limited implementation plan should be followed.
The specification should be developed and processed through the Specification Committee.
6. 1.2 1995 Constrl.lction Season
Step 1: If any disputes arise during the implementation, the maximum of 20% natural fine aggregates should be used as a referee.
The implementation plan for this specification appears very short. However, the testis very
simple and the device is not expensive. Furthermore, there has been a great need expressed
by both the contractors and CDOT to quantify the fine aggregate angularity.
24
When the SHRP gyratory arrives, the angularity specification recommended by SHRP should be
evaluated with the new gradation requirements and N'n".
6.2 Draft Specification
Currently the Standard Specifications, Section 703.04, limits the maximum natural fine aggregates
that are allowed. This should be deleted. Table 403-1 in the project special provisions should
have a new line added that states "Uncompacted Voids (%), AASHTO TP3", with a "fill-in-the
blank". The Design Manual should provide the guidance for the Materials Engineers to fill-in-the
blank. The guidance should be that shown in Table 6.
6.3 Other Modifications Necessary
The mix design sheets currently include the bulk specific gravity of the combined aggregate blend
(G,.). This value should still be reported. The G' b of the fine aggregate blend should be
reported in addition. This will provide the information necessary for monitoring the uncompacted
voids during construction.
The fine aggregates in recycled asphalt pavement (RAP) have been extremely rounded in some
cases. When RAP is used, the fine aggregates should be extracted from the RAP and tested
in the AASHTO TP3 procedure. The fine aggregates in the RAP should be blended with the
other fine aggregates in the percentages used in the mixture. In the Central Laboratory, the
effective specific gravity of the fine aggregates (G .. ) should be used for the RAP . . In the Region
Laboratories, a solvent wash with biodegradable solvents will be required to obtain extracted fine
aggregates from the RAP for testing the angUlarity.
A CP-L should be written to replace AASHTO TP3. Methods Band C need to be eliminated.
The provisions to account for RAP need to be added. After the distribution of the equipment and
training, any procedural clarifications identified should be added.
25
7.0 Conclusions
1) The French P 18-564 and MSHTO TP3 methods produre very comparable results. Using
MSHTO TP3, a minimum uncompacted voids of 45.0% would include ali of the manufactured
fine aggregates, and a value of 46.0% would exclude most of the natural fine aggregates. Using
the French P 18-564, a flow time of 33.0 to 35.0 seconds delineated between all of the natural
and manufactured fine aggregates.
2) AASHTO TP3 is recommended for use as a specification. The specification should be
applied to the tested blend of fine aggregates. The blend should not be calculated from the
values of the individual components. However, if all individual stockpiles exceed the
specification, . testing on the blend is not required.
3) Specification values should be those shown in Table 6. A specification of 46.0% allows
natural fine aggregate in a similar quantity that is currently allowed. A specification of 45.0% will
likely allow more natural fine aggregates. These values are higher than the SHRP
recommendations but lower than the French recommendations.
4) The proposed specification should be implemented in a staged and methodical manner.
26
8.0 References
1. Asphalt Concrete Mix Design and Field Control. Technical Advisory T 5040.27, Federal Highway Administration, U.S. Department of Transportation, March 10, 1988,27 pages.
2. Tentative Method of Test for Uncompacted Void Content of Fine Aggregate. National Sand and Gravel Association and National Ready-Mix Concrete Association, Washingion, D.C., (Undated)
3. Test Method for Measuring Fine Aggregate Angularity. Michigan Test Method 118-90, Approved November, 1990.
4. Determining the Flow Coefficients of Sand. Tentative Standard P 18-564, February 1981.
5. Kandhal, P.S., J.B. Motter, and M.A. Khatar (1991), "Evaluation of Particle Shape and Texture: Manufactured Versus Natural Sands," Transportation Research Record 1301, Transportation Research Board, National Research Council, Washington, D.C., pp. 48-56.
6. Mogawer, W.S., and K.D. Stuart (1992), Evaluation of Test Methods Used to Quantify Sand Shape and Texture," Transportation Research Record 1362, Transportation Research Board, National Research Council, Washington, D.C., pp. 28-37.
7. Aschenbrener, T.B. (1992), "Comparison of Colorado Component Hot Mix Asphalt Materials With Some European Specifications," Colorado Department of Transportation, CDOTDTD-R-92-14, 65 pages.
8. SUPERPAVE Asphalt Mixture Design, February 1994, National Asphalt Training Center, Demonstration Project 101, FHWA Office of Technology Applications, Washington, D.C., and the Asphalt Institute Research Center, Lexington, Kentucky, 147 pages.
9. Aschenbrener, T.B. (1993), "Determining Optimum Asphalt Content With the Texas Gyratory Compactor," Colorado Department of Transportation, CDOT-DTD-R-93-23, 78 pages.
27
Appendix A
AASHTOTP3
~.
June, 1993 Passed committee C-9 Letter Ballot Several Editorial corrections Made
Standard Test Method for Uncompacted Void Content of Fine Aggregate (As Influenced by Particle Shape, Surface Texture, and Grading)
1. SCOPE
1.1 This method describes the determination of the loose uncompacted void content of a sample of fine aggregate. When measured on any aggregate of a known grading, void content provides an indication of that aggregate's angularity, sphericity, and surface texture compared with other fine aggregates tested in the same grading. When void content is measured on an as-received fine aggregate grading, it can be ari indicator of the effect of the fine aggregate on the workability of a- mixture in which it may be used.
1.2 Three procedures are included for the measurement of void content. Two use graded fine aggregate (standard grading or -as-received grading), and the other uses several individual size fractions for void content determinations:
1. 2.1
1.2 . 2
1. 2 . 3
1. 2.4
Standard Graded Sample (Method -A) This method uses a standard fine _- aggregate grading that is obtained by combining individual sieve fractions from a typical fine aggregate sieve analysis. See the section on preparation of Test Samples for the grading.
Individual Size Fractions (Method B) -- This method uses each of three fine aggregate size fractions: (a) 2. 36-mm (No.8) to 1. 18-mm (No. 16); (b) 1.18-mm (No. 16) to 600-Jlm (No. 30); and (c) 600-Jlm (No. 30) to 300-Jlm (No. 50) . For this method, each size is tested separately.
As-Received Grading (Method C) -- This method uses that portion of the fine aggregate finer than a 4.75-mm (No.4) sieve.
See the section on Significance --and Use for guidance on the method to be used.
1 . 3 The values stated in SI Units shall be regarded as the standard.
1.4 This standard does not purport to address all of the safety problems, if any, associated with its use. It is
Al
the responsibility of the user of this . standard to establish' appropriate safety and health practices and determine the applicability of regulatory limitations prior to use.
2 . REFERENCED DOCUMENTS
2.1 ASTM Standards
B 88 - Specification for Seamless Copper Water Tube.
B 88M - Specification for Seamless Copper Water Tube (Metric)
C 29/29M - Test Method for unit Weight and Voids in Aggregate
C 117 Test Method for Materials Finer than 75-~m (No. 200) Sieve in Mineral Aggregates by Washing.
C 125 Terminology Relating to Concrete and Concrete Aggregates.
C 128 Test Method for specific Gravity and Absorption of Fine Aggregate.
C 136 Test Method for Sieve Analysis of Fine and Coarse Aggregates.
C 702 Practice for Reducing Field Samples of Aggregate to Testing Size.
C 778 Specification for Standard Sand
D 75 Practice for Sampling Aggregates.
2.2 ACI Document
ACI 116R Cement and Concrete Terminology'
3 . TERMINOLOGY
3.1 Terms used in this standard are defined in Terminology C 125 or ACI 116R.
Copies may be obtained from the American Concrete Institute, Box 19150, Detroit, MI 48219 .
A2
-3-
4. SUMMARY OF TEST METHOD
4.1 A nominal 100-mL calibrated cylindrical measure is filled with fine aggregate of prescribed grading by allowing the sample to flow through a funnel from a fixed height into the measure. The fine aggregate is struck off, and its mass is determined by weighing. Uncompacted void content is calculated as the difference between the volume of the cylindrical measure and the absolute volume of the fine aggregate collected in the measure. Uncompacted void content is calculated using the bulk dry specific gravity of the fine aggregate. Two runs are made on each sample and the results are averaged.
4.1.1
4.1. 2
For a graded sample (Method A or Method C) the percent void content is determined directly, and the average value from 'two runs is reported.
For the individual size fractions (Method B), the mean percent void content ·is calculated using the results from tests of each of the three individual size fractions.
5. SIGNIFICANCE AND USE
5.1 Methods A and B provide percent void content determined under standardized conditions which depends on the particle shape and texture of a fine aggregate. An increase in void content by these procedures indicates greater angularity, les's sphericity, or -rougher surface texture, or some combination of the three factors . A decrease in void content results is associated with more rounded, .spherical, smooth surfaced fine aggregate, or a combination of these factors.
5.2 Method C measures the uncompacted void content of the minus 4.75-mm (No.4) portion of the as-received material. This void content depends on grading as well as particle shape and texture.
5.3 The void content determined on the standard graded sample (Method A) is not directly comparable with the average void content of the three individual size fractions from the same sample tested separately (Method B» A sample consisting of single size particles will have a higher void content than a graded sample. Therefore, use either one method or the other as a comparative measure of shape and texture, and identify which method has been used to obtain the reported data. Method C does not provide an indication of shape and texture directly if the grading from sample to sample changes.
A3
5 . 3 . 1
. 5 . 3.2
5.3;3
5.3 .4
-4-
The standard graded sample (Method A) is most useful as a quick test which indicates the particle shape properties of a graded fine aggregate. Typically, the material used to make up the standard graded sample can be obtained from the remaining size fractions after performing a single sieve analysis of the fine aggregate.
Obtaining and testing individual size fractions (Method B) is more time consuming and requires a larger initial sample than using . the graded sample. However, Method B provides additional information concerning the shape and texture characteristics of individual sizes.
Testing samples in the as-received grading (Method C) may be useful in selecting proportions of components used in a variety of m:i,xtures. In general, high - void content suggests that the material could be improved by providing additional fines in the fine aggregate or more cementitious material may be needed to fill voids between particles. -
The bulk dry specific gravity o-f the fine aggregate · is used ~n calculating the void content. The effectiveness of these methods of determining void content and its relationship to particle shape' and texture depends on the bulk specific gravity of the various size fractions being equal, or nearly so. The void content is actually a function of the volume of each size fraction. If the type of rock or minerals, or its porosity, in any of the size fractions varies markedly it may be necessary to determine the specific gravi ty of the size fractions used in the test.
5.4 Void content information from Methods A, B, or C will be useful as an indicator of properties such as: the mixing water demand of hydraulic cement concrete; flowability, pumpability, or workability factors when " formulating grouts or mortars; or, in bituminous concrete, the effect of the fine aggregate on stability and voids in the mineral aggregate; or the stability of the fine aggregate portion of a base, course aggregate -.
A4
-5-
6 . APPARATUS
~.1 CYlindrical Measure -- A right cylinder of approximately 100-mL capacity having an inside diameter of approximately 39-mm and an inside height of approximately 86-mm made of drawn copper water tube meeting ASTM Specification B 88 Type M, or B 88M Type C. " The bottom of the measure shall be metal at least 6-mm thick, shal l be firmly sealed to the tubing, and shall be provided with means for aligning "the axis of the cylinder with that of the funnel. See Figure 1.
6.2 Funnel -- The lateral surface of the right frustum of a cone sloped 60 ± 4· from the horizontal with an opening of 12.7 ± 0.6-nun diameter. The funnel section shall be a piece of metal, smooth on "the inside and at least 38-mm high. It shall haVe a volume of at least 200-mL or shall be provided with a supplemental glass or metal container to provide the required volume. See Figure 2.
Note 1 -- Pycnometer top C9455 sold by Hogentogler and Co., Inc., 9515 Gerwig, Columbia, Maryland 21045, 410-381-2390 is satisfactory for the funnel section, except that the size of the opening has to be enlarged and any burrs or lips that are apparent should be removed by light filing or sanding before use. This pycnometer top must be used with a suitable glass jar with the bottom removed (Figure 2) .
6.3 Funnel stand -- A three or four legged support capable of holding the funnel firmly in position with the axis of the funnel colinear (within a 4· angle and a displacement of 2 rom) with the axis of the cylindrical measure. The funnel opening shall be 115 ± 2 mm above the top of the cylinder. A suitable arrangement is shown in Figure 2.
6.4 Glass Plate -7 A square glass plate approximately 60 mm by 60 mm with a minimum 4-mm thickness used to calibrate the cylindrical measure.
6.5 Pan -- A metal or plastic pan of sufficient size to contain the funnel stand and to prevent loss of material. The purpose of the pan is to catch and retain fine aggregate particles that overflow the measure during filling and strike off.
6 . 6 Metal spatula with a blade approximately 100-mm long, and at least 20-mm wide, with straight edges. The end shal l be cut at aright angle to the edges. The straight edge of the spatula blade is used to strike off the fine aggregate.
AS
7 .
-6-
6.7 Scale or balance accurate and readable tot 0.1 g within the range' of use, capable of weighing the cylindrical measure and its contents.
SAMPLING
7.1 The sample(s) used for this test shall be obtained using Practice D 75 and Practice C 702, or from sieve analysis samples used for Test Method C 136, or from aggregate extracted from a bituminous concr.ete specimen. For Methods A and B, the sample is washed over a l50'-jlm (No. 100) or 75-jlm (No. 200) .sieve in accordance with Test Method C 117 and then dried and sieved into separate size fractions according to Test Method C 136 procedures. Maintain the necessary size fractions obtained from one (or more) sieve analysis in a dry condition in separate containers for each size. For Method C, dry a split of the as-received sample in accordance with the drying procedure in Test Method C 136. .
8. CALIBRATION OF CYLINDRICAL MEASURE
8.1 Apply a light coat of grease to the top edge of the dry, empty cylindrical measure. Weigh the measure, grease, and glass plate. Fill the measure with freshly boiled, deionized water at a temperature of 18 to 24" C. Record the temperature of the water. Place the glass 'plate on the measure, being sure that no air bubbles remain. Dry the outer surfaces of the measure and determine the combined mass of measure, glass plate, grease, and water by weighing. Following the final weighing, remove the grease, and determine the mass of· the clean, dry, empty measure for subsequent tests.
8.2 Calculate the volume of the measure as follows:
v = 1000 M D where:
v = volume of cylinder, mL
M = net mass of water, g
D = density of water (see table in C 29/C 29M for density at the temperature used), Kg/m3
Determine the volume to the nearest 0 . 1 mL.
Note 2 -- If the volume of the measure is greater than 100.0 mL, it may be desirable to grind the upper edge of the cylirider until the volume is exactly 100.0 mL, to simplify subsequent calculations .
A6
- 7-
9 . PREPARATION OF TEST SAMPLES
9. 1 Method A Standard Graded Sample -- weigh out and combine the following quantities of fine aggregate which has been dried and sieved in accordance with Test Method C 136.
Individual Size ,Fraction Mass, g
2. 36-rom (No. 8) to 1. 18-rom (No. 16) 44
1. 18-mm (No. 16) to 600-jlm (No. 30) 57
600-/Lm (No . 30) to 300-jlm (No. 50) 72
300-/Lm (No. 50) to 150-/Lm (No. 100) 17
190
The tolerance on each of these amounts is ± 0. 2 g.
9 . 2 Method B Individual Size Fractions Prepare a separate 190-g sample of fine aggregate, dried and sieved in accordance with Test Method C136, for each of the following size fractions: '
Individual Size Fraction Mass. g
2.36-mm (No.8) to 1.18-mm (No. 16)
1.18-mm (No. 16) to 600-/Lm (No. 30)
600-/Lm (No. 30) to 300-/Lm (No. 50)
The tolerance on each of these amounts is ± 1 g. mix t hese samples together. Each size is separately.
190
190
190
Do not tested
9 . 3 Method C - As Received Grading -- Pass the sample (dried in accordance with Method C 136) through a 4.75-rom (No. 4) sieve. Obtain a 190 ± 1-g sample of the material passing the 4.75-rom (No.4) sieve for test.
9 . 4 Specific Gravity of Fine Aggregate - If the bulk dry specific gravity of fine aggregate from the' source is unknown, determine it on the minus 4. 75-mm (No.4 ) material according to Test Method C 128. Use this value in subsequent calculations unless some size fractions differ by more than 0.05 from the specific gravity typical of the complete sample, in which case the specific gravity of the fraction (or fractions) being tested must be determined. An indicator of differences in specific gravity of various particle sizes is a comparison of specific gravities run on the fine
A7
-8-
aggregate in different gradings. Specific gravity can be run on gradings with and without specific size fractions of inte.rest. If specific gravity differences exceed 0.05, determine the specific gravity of the individual 2.36.-mm (No.8) to ISO-11m (No. 100) sizes for use with Method A or .the individual size fractions for use with Method B either by direct measurement or by calculation using the specific gravity data on gradings with and without the size fraction of interest. A difference in specific gravity of 0.05 will change the calculated void content about one percent. .
10. PROCEDURE
10 .1 Mix each test sample with the spatula until it appears to be homogeneous. Position the jar and funnel section in the stand and center the cylindrical measure as shown in Figure 2. Use a finger to block the opening of the funnel. Pour the test sample into the funnel. Level the material in the funnel with the spatula. Remove the finger and allow the sample to fall freely into the cylindrical measure.
10.2 After the funnel empties, strike-off excess heaped fine aggregate from the cylindrical measure by a single pass of the spatula with the width of the blade vertical using the straight part of its edge i n light contact with the. top of the measure. Until this operation is complete, exercise care to avoid vibr~tion or any disturbance that could ·cause compaction of the fine aggregate in the cylindrical measure. (Note 3) Brush adhering grains from the outside of the container and determine the mass of the cylindrical measure and contents to the nearest 0.1 g. Retain all fine aggregate particles for a second test run.
Note 3 -- After strike-off, the cylindrical measure may be tapped lightly to compact the sample to make it easier to transfer the container to scale or balance without spilling any of the sample.
10 . 3 Recombine the sample from the retaining pan and cylindrical measure and repeat the procedure. The results of two runs are averaged. See the Calculation section.
10 . 4 Record the mass of the empty measure. Also, for each run, record the mass of the measure and fine . aggregate.
A8
-9-
ll . CALCULATION
ll.l Calculate the qncompacted voids for each determination as follows:
U = V - (FIG) x lOO V
V = volume of cylindrical measure, mL
F = net mass, g, of fine aggregate in measure (Gross mass minus the mass of the empty measure).
G = bulk dry specific gravity of fine aggregate.
U = uncompacted voids, percent, in the material.
ll . 2 For the Standard Graded Sample (Method A)· calculate the average uncompacted voids for the two determinations and report the result as Us.
ll.3 For the Individual Size Fractions (Method B) calculate:
lL 3.1
1L3 . 2
First, the average uncompacted voids for the determinations made on each of the three sizefraction samples:
U, = Uncompacted Voids, 2.36-mrn (No . 8 ) to Ll8-nun (No. 16), percent
U2 = Uncompacted Voids·, L 18-mrn (No. l6) to 600-~m (No. 30), percent
U3 = Uncompacted Voids, 600-~m (No . 30 ) to 300-~m (No. 50), percent
Second, the mean uncompacted voids (Um
)
including the results for all three sizes:
Um = (U, + U2 + U3) / 3
ll.4 For the As-Received grading (Method C) calculate the average uncompacted voids for t he two determinations and report the result as UR •
12 . REPORT
12. 1 For the Standard Graded Sample (Method A) report:
12 . 1. 1
12 . 1.2
The Uncompacted Voids (Us) in percent to the nearest one-tenth of a percent (0.11 ) .
The specific gravity value us ed in the calculations .
A9
-10-
12.2 For the Individual Size Fractions (Method 8) report the following percent voids to the nearest one-tenth of a percent (0.1%):
12.2.1
12.2.2
12.2.3
Uncompacted Voids fo·r size fractions: (a) 2.36-mm (No.8) to 1.18-nun (No. 16) (U,); (b) 1.18 mm (No. 16) to 600-llm (No. 30) (U2); and (c) 600-llm (No. 30) to 300-llm (No. 50) (U3 ).
Mean Uncompacted Voids (Um).
specific gravity value(s) used in the calculations, and whether the specific gravity value (s) were determined on a graded sample o·r the individual size fractions used in the test.
12.3 For the As-Received Sample (Method C) report:
12.3.1
12.3.2
The uncompacted voids (UR) in percent to the nearest one-tenth of a percent (0.1%) .
The ·specific gravity value used in the calculation .
13 . PRECISION AND BIAS
13; 1 Precision
13.1.1.
13 . 1.2
13 . 1 . 3
The single-operator standard deviation has been found to be 0.13 percent voids (IS), using the graded standard silica sand as described in Specification C 778. Therefore, results of two properly conducted tests by the same operator on similar samples should not differ by more than 0.37 percent (025).
The multilaboratory standard deviation has been found to be 0.33 percent (IS) using the standard fine aggregate as described in Specification C 778. Therefore, results of two properly conducted tests by different laboratories on similar samples should not differ by more than 0.93 percent (02S).
The above statements pertain to void contents determined on "graded standard sand" as described in Specification C 778, which is considered rounded, and is graded from 600-IlID (No. 30) to 150-llm (No. 100), and may not be typical of other fine aggregates. Additional precision data are needed for tests of fine aggregates having different levels. of
AlO
13 . 2 Bias
-11-
angularity and texture tested in accordance with this Test Method .
since there is no accented reference material suitable for determining the bias for the procedures in this Test Method, bias has not been determined .
14 . Keywords: Angularity, Fine Aggregate, particle Shape, Sand, Surface Texture, Void Content
All
I 41 mm A:rox. I
. 1---1. ~39::....:..:.:.:mm:.:...:...IAp=-=:prox=---. ----;, '
Copper Pipe
~
...----epoxy Resin Filled Join~
I ·i
3.5
>< e « E E CD co
±O.4 mm
I f
Metal Plate / \pprox. 6 mm dia. Drilled Hole Approx. 3 mm Deep . Minimum 6 mm Thick
Used for Centering Container on a Mating Dowel in the Center of the Stand Base
Figure 1: Nominal 1 00 mL Cylindrical Measure
A12
I L
Pan to Retain Fine Aggregate
Particles
r-
I
-
;1ne Quart (Uter) Mason Jar Pycnometer (Remove bottom of glass jar)
I ~ I F-h-J------',-i , rl--------~~
..-J Min. 38 mm~; __ L : 60o ±4°
-/i'f 12.7±0.6 mm dia !
n
I 11S±2 mm , i
Nominal 100 mL Melsur~ ~~ -
I , I !
I j ,
Section Through Center of Apparatus
Figure 2: Suitable Funnel Stand Apparatus with Cylindrical Measure in Place
A13
Appendix B
French Tentative Standard P 18-564
"
'.
• ¥ = i • .. ~
::::; . • ~! J •
=~ ::1 =2 < 3
• =-i < • .~
:.6
~ · • : · .. · · i ~ -~ ; :
:-l"1 ;t ~, --
"
., .... '"' ... --=
! ~ • :::l ~
-= ~ § .... ., • ..! ,
u • 3-· ... 3
-0 z: ~
~
= .,. 5 z -! ~ .;;
• • ... ;;
~ < :.. ~ ~
" ~
!
I I I
I I I I
I
:ilA.~DA.'lD
AGG<ti:G"E5
:>ET~!Uml[~G THE f LO,/
COEfF[C[L~l Of S~~D
-z.( Q
.0 .J -.....
fOREIIO!D' :.:.~ .....
~, .. <;..
"nOil has issued cn~q.~~d~·r,;~"·~{~i;;f';'it th .. status of ... G r ' . '... . ...
a tent~cive stan~Yd. ci" bejng rtac·eJs~.ry to con(irm. by
I!xpericnce I tn<~&Lue: oi~e'<~'c. .• :';
~.
\>" Afte.r ~ period of thre~ ye~rs thi, stmdard \liLl be
re.-exUtined and ies st.acus m.ay be ch~,ed.
[. SCOP~
2.
The. purpose of chis tentative standard i.l to specify
the procedure for me.asuring the flow coefficient of
s~d.
f[ELD Of APPL[CATION
Th is standard applies to smds (U of n&cural or artificiaL.
origin. used for roadvays .
3. DEf[~nnON
Flow coefficient for .a smd time: of flo,", I expressed in
second •• for lOOO I of I;hi' s.Jnd to pAIS throuah a. iiven
orifice 4
(1) See standard N P18'-·lOl "Aggrex_tes - !erminolo&y, definitions. classifications" in prep.Jration
Cocnme:nt.J reLlting to chis tentative .Jt~d.ard I are to be sent to AYNOR, TOUR EUROPE CEDEX 7,
i 92080 ?AR[S l.A DEfENSE. I I
GRA.'1Ul.ATS • DETERNIN"TiON DU cnUnCU:NT D' ECOULE~Nl DES' SADLES
81
I
I , , i I
P I8 -56!.
' . .';'PPARATU5
rill!! app~r.cu.5 used for chis flow C~st (see figure L,. comprises :
[VO funnels 1 83 an high and made from poLycarbonate •
• :me wi~h a 12 c;m opening. th~ ocher with a 16 am opening
a cylindric~L body which. can be fitted onco ,Jne ,Jr the
Ilchl!r !ole the funneLs, inside diameter 90 atm., outside
diamHer 100 IIUIl and 125 _ hiSh,
.a scand , provided vitti .z cover pLate whieh. Can move on a
pi.vot to open 'and close the orifice . I
a container of sufficient c.pacity to receive the material
flaving from the funnel.
Si.eves. 0.08 - 2 and £.. ram. ,
a chr-onomecer recording to 1/10 seconds ,
a b.lmce, minimum ccpacity 2 kl. capable of weig~ling to
uiehin 1 g .
an oven, capable of reachina lOSoe.
5 . ~IATERIAL fOR TESn:;C
The sample shall be prepared l.n accordmce with standard
~f P 18-55) (lJ.
The \.Ieighc of the test sample ;n.all be .such th.t the p.art
dried ole lOSoC and p.a.ssing chreugh the 2 or 4 IDItL sieve L.5
1000 II ~ I I ·
6 . CARRYING OllT THE TEST
Select .t funnel, using the 12 mID openin& for 0/2 sand or
16 mm opening for 0/4. sand , md pl.ac.e it on itt stand, its
opening closed off . Pl.ace the test s.uaple in the. funnel and
restrict the height of fall Co avoid compacting the macerial.
(lJ )If PI8-55) "Aggrell~te. - PrepOlrin, ~ .~",pl. for testinlt"
B2
, P I B - ;64
Open the orifice.and at the s.u-te ti.me st.rt the chr,Jnarrec~r :
determine the c.ime.. Co wi~hin 1/10 second, for .t·Ll the
:'i1.1cet'i&l to flow through the funnel. Repeat the test five
time3. uain& the S.aIDe cest sa.m.pLe .
7. ~XPitESSt/iC IKE R.ESULIS
The: flow cot: ffic:ienc EC (0/0) . ex.pressed in sec:onq,s . of che
sand !)eing tested, i.s the aver"'Ce of che five measurements.
rounded off to ehe nearest second .
.. -
133
o i:nens ions
in
mi i limecres
OJ
i , ,
.nl CQI
.'
p 18 • jo~
- - - - - -t--"'-':'.!.:..;
- Funne 1
t ~116 1<
Cover .place
Con r. .line r
, .
B4